基于LST-EVI特征空间的土壤水分含量反演

干旱是人类历史上的重大自然灾害之一,而土壤水分是干旱监测最重要的指标。利用遥感手段反演地表土壤水分,可以充分反映土壤水分的时空变化特征,适合进行大范围动态监测。研究基于Landsat TM数据,运用普适性单通道算法得到地表温度(LST,Land Surface Temperature),然后选用增强型植被指数(EVI,Enhanced Vegetation Index),构建了LST\|EVI特征空间,计算出温度植被干旱指数(TVDI,Temperature\|Vegetation Dryness Index)。在对实测土壤含水量数据和对应TVDI值进行回归分析的基础上,反演出2010年6月14日黄骅市自然地表20 cm深度处的体积含水量。结果表明:TVDI方法在该研究区是完全可行的,拟合精度较高;研究区自然地表土壤体积含水量分布差异明显,中等含水量地区面积最大,西南和部分北部地区含水量较低,而含水量高的区域主要分布在苇洼和沿海地区。

Soil Moisture Estimation based on the LST-EVI Feature Space

Drought is one of the most catastrophic natural disasters in human history.Soil moisture is the key parameter in study of drought monitoring.Application of remote sensing techniques for estimation of surface soil moisture can adequately reveal the spatial and temporal variations,which is suitable for a large scale dynamic monitoring.Based on Landsat TM data,land surface temperature LST was derived by means of universal single\|channel algorithm.EVI was acquired to build the LST\|EVI feature space,and calculated the temperature vegetation dryness index(TVDI).TVDI and observed soil moisture data were analyzed and a linear curve was fitted to estimate volumetric soil moisture content in the depth of 20cm for June 14th,2010,in areas of natural surface in Huanghua City.TVDI is feasible for monitoring soil moisture content of the study area.Spatial distribution of volumetric soil moisture content varied significantly.Area of moderate soil moisture was the largest.Areas with critically low soil moisture content mostly were upland fields,which is located in the southwest and part of the north region.However,the areas with high soil moisture content mainly located in the reed swamp and coastal bare land.